RED-Metrics-Tracker

Simple RED Metrics tracker able to instrument flask views using prometheus metrics.

Install

With pip, of course:

pip install red-metrics-tracker

Instrumenting

Tracking all methods and exceptions for all requests on a view:

app = flask.Flask(__name__)

@app.route("/endpoint")
@FlaskRedMetricsTracker.track()
def do_things():
     ...
     return "OK"

Limiting metrics tracking to specific methods:

@app.route("/endpoint", methods=["POST", "GET", "PUT", "DELETE"])
@FlaskRedMetricsTracker.track(methods=["POST", "DELETE"])
def do_things():
     ...
     return "OK"

Limiting to specific exceptions:

@app.route("/endpoint/<condition>")
@FlaskRedMetricsTracker.track(exceptions=MySpecialException)
def do_things():
     ...
     if condition == "throw":
        raise MySpecialException
     return "OK"

Filters may be combined, of course:

@app.route("/endpoint/<condition>", methods=["POST", "GET", "DELETE"])
@FlaskRedMetricsTracker.track(methods=["GET", "POST"]exceptions=MySpecialException)
def do_things():
     ...
     if condition == "throw":
        raise MySpecialException
     return "OK"

Exposition

Feel free to use prometheus_client to run an http_server serving a /metrics endpoint at its configured port.

For the people who’d like the /metrics endpoint to be part of their flask app, there’s a blueprint for that:

import flask
from red import metrics_blueprint

app = flask.Flask(__name__)
app.register_blueprint(metrics_blueprint)

Metrics are now available at you app’s url under the /metrics endpoint.